2004 : WHAT'S YOUR LAW?

[ print ]

Cognitive neuroscientist; Author, Optical Illusions: The Science of Visual Perception
Seckel's Law

Seckel's First Law

Visual Perception is Essentially an Ambiguity Solving Process.

Most of us take vision for granted. After all, it comes to us so easily. With normal vision we are able to navigate quickly and efficiently through a visually rich three-dimensional world of light, shading, texture, and color—a complex world in motion, with objects of different sizes at differing distances. Looking about we have a definite sense of the "real world".

In fact, our visual system is so successful at building an accurate representation of the real world (our perception) that most of us do not realize what a difficult task our brain is performing. Without conscious thought, our visual system gathers and interprets complex information, providing us with a seamless perception of our environment. The complexities of how we perceive are cleverly concealed by a successful visual system.

It might seem reasonable for us to assume that there is a one-to-one mapping between the real world and what you perceive—that your visual system "sees" the retinal image, in much the way that a digital camera records what it "sees."

Although it seems like a useful analogy, there is no real comparison between our visual system and a camera beyond a strictly surface level. Furthermore, this comparison trivializes the accomplishments of our visual system. This is because a camera records incoming information, but our brain interprets incoming information. Furthermore, it feels to us as if a photograph reproduces a three-dimensional world, but it doesn't. It onlysuggests one. The same visual system that interprets the world around us also interprets the photograph to make it appear as a three-dimensional scene.

Our perceptions are not always perfect. Sometimes our brain will interpret a static image on the retina in more than one way. A skeleton cube, known as a Necker cube, is a classic example of a single image that is interpreted in more than one way. If you fixate on this cube for any length of time, it will spontaneously reverse in depth, even though the image on the retina remains constant. Our brain interprets this image differently because of conflicting depth cues.

The great 19th century German physicist and physiologist Hermann Von Helmholtz first discovered the basic problem of perception over one hundred years ago. He correctly reasoned that the visual information from our world that is projected onto the back of the retina is spatially ambiguous. Helmholtz reasoned that there can be an infinite variety of shapes that can give rise to the same retinal image, as long as they subtend the same visual angle to the eye.

However, the concept of visual ambiguity is far deeper than what Helmholtz originally proposed, because it turns out that any one aspect of visual information, such as brightness, color, motion, etc, could have arisen from infinitely many different conditions. It is very hard to appreciate this fact at first, because what we perceive in a normal viewing environment is not at all ambiguous.

If all visual stimuli are inherently ambiguous, how does our visual/perceptual system discard the infinite variety of possible conditions to settle on the correct interpretation almost all the time, and in such a quick and efficient manner? The problem basically stated is, how does the visual system "retrieve" all of the visual information about the 3D world from the very limited information contained in the 2D retinal image? This is a basic and central question of perception.

Studying the visual system only at one level will never result in a full understanding of visual perception. Many of the underlying mechanisms that mediate vision may be even "messier" than previously thought, with cross-feedback from more than one level of visual processing contributing to processing at another level. UCSD vision scientist V.S. Ramachandran is correct when he believes that it is time to "open the black box in order to study the responses of nerve cells," but he is also probably right to promote his Utilitarian Theory of Perception, which argues for a clever "bag of tricks" that the human visual system has evolved over millions of years of evolution to resolve the inherent ambiguities in the visual image. Visual perception is largely an ambiguity-solving process.

The task of vision scientists, therefore, is to uncover these hidden and underlying constraints, rather than to attribute to the visual system a degree of simplicity that it simply does not possess.

Seckel's Second Law

Our Visual/Perceptual System is Highly Constrained.

Sometimes our perceptions are wrong. Often these errors have been classified as illusions, dismissed by many as failures of the visual system, quirky exceptions to normal vision.

If illusions are not failures of the visual system, then, what are they? After all, we do categorize a number of different perceptual experiences as "illusions". What makes them fundamentally different than those we perceive as normal?

One difference is a noticeable split between your perception and conception. With an illusion, your perception is fooled but your conception is correct—you're seeing something wrong (your mis perception), but youknow it's wrong (your correct conception). Initially, your conception may be fooled too, but at that point you are unaware that you are encountering an illusion. It is only when your conception is at odds with your perception that you are aware that you have encountered an illusion.

Furthermore, in almost all pictorial illusions (where the meaning of the image is not ambiguous), your perceptions will continue to be fooled, even though your conception is fine, no matter how many times you view the illusion. It does not matter how old you are, how smart you are, how cultured you are, or how artistic you are, you will continue to be fooled by these illusions over and over again. In fact, you cannot "undo" your incorrect perceptions, even with extended experiences, worldly knowledge, or training. It is more important for your visual system to adhere to these constraints than to violate them because it has encountered something unusual, inconsistent, or paradoxical. This indicates that your visual/perceptual system is highly constrained on how it interprets the world.

It is not my intention to cause the reader to think that visual perception is unreliable and untrustworthy. This would be a mistake as, for the most part, our perceptions of the world are veridical. However, how we perceive the world is not a mirror image of reality, but an actively and intelligently constructed one that allows us to have the best chances for survival in a complicated environment.